含AI模型机电系统环境适应性试验与评估方法

doi:10.3901/JME.2025.16.358

机械工程学报 ›› 2025, Vol. 61 ›› Issue (16): 358-373.doi: 10.3901/JME.2025.16.358

• 交叉与前沿 • 上一篇

扫码分享

含AI模型机电系统环境适应性试验与评估方法

张书锋^1,2, 宋国锋^1,2,3, 李兴格⁴, 陈循^1,2

1. 国防科技大学智能科学学院长沙 410073；
2. 国防科技大学装备状态感知与敏捷保障全国重点实验室长沙 410073；
3. 西北机电工程研究所咸阳 712099；
4. 军事科学院国防科技创新研究院北京 100071

接受日期:2024-09-05 出版日期:2025-02-23 发布日期:2025-02-23
作者简介:张书锋，男，1987年出生，博士，副研究员，硕士研究生导师。主要研究方向为智能装备综合保障。E-mail：sfzhang@nudt.edu.cn;李兴格(通信作者)，男，1995年出生，博士，助理研究员。主要研究方向为天基无人系统健康管理。E-mail：lixingge1995@126.com
基金资助:
国防科工局国防基础科研计划重点(WDZC20205500303)和湖南省研究生科研创新(CX20210061)资助项目

Environmental Worthiness Testing and Evaluation of Electromechanical System Containing AI Model

ZHANG Shufeng^1,2, SONG Guofeng^1,2,3, LI Xingge⁴, CHEN Xun^1,2

1. College of Intelligent Science and Technology, National University of Defense Technology, Changsha 410073;
2. National Key Laboratory of Equipment State Sensing and Smar Suppot, National University of Defense Technology, Changsha 410073;
3. Northwest Institute of Mechanical and Electrical Engineering, Xianyang 712099;
4. Defense Innovation Institute, Chinese Academy of Military Science, Beijing 100071

Accepted:2024-09-05 Online:2025-02-23 Published:2025-02-23

摘要/Abstract

摘要： 机械化、信息化和智能化的深度融合是未来装备的发展趋势，越来越多的人工智能(Artificial intelligence, AI)模型嵌入机电系统，其环境适应性评估不仅需要考虑硬件失效，更需要考虑环境对AI模型功能的影响。针对含AI模型机电系统的环境适应性评估需求，以数据分布度量函数为切入点，讨论含AI模型机电系统的环境适应性内涵，构建基于数据分布度量的环境适应度评价指标。考虑到基于性能数据的建模方法可以提供更多定量化信息，能够将模型性能变化数据与环境适应度函数结合起来，分别提出基于性能变化轨迹和性能变化量分布的环境适应性定量评估方法。最后，以含AI模型视觉感知系统为例，选择浓雾恶劣天气和振动两种环境应力开展试验，获取视觉感知系统元数据集在不同测试环境样本下的性能变化量，得到系统的环境适应度曲线和平均失效差异值，验证方法的准确性与有效性。

关键词: 人工智能模型(AI模型), 环境适应性, 数据分布度量, 性能数据建模, 元数据集

Abstract: The profound integration of mechanization, informatization, and intelligentization represents the development trend of future equipment. An increasing number of electromechanical systems embed artificial intelligence(AI) models, necessitating the evaluation of environmental worthiness. This evaluation must not only consider hardware failures but also take into account the impact of the environment on the functionality of AI models. Aiming at the requirements of environmental worthiness evaluation for electromechanical systems containing AI models(ESAM), it takes the data distribution measurement function as a starting point, discusses the connotation of environmental worthiness in ESAM, and constructs an environmental worthiness index based on data distribution measurement. A quantitative evaluation method of environmental worthiness based on performance trajectories and distribution is proposed, considering that modeling methods using performance data can provide more quantitative insights by integrating model performance with environmental worthiness functions. Finally, using an AI-based visual perception system as an example, the study examines performance data under varying fog and vibration conditions in testing environments. It derives environmental worthiness curves and average failure discrepancy values, confirming the validity and effectiveness of the proposed method.

Key words: artificial intelligence model(AI), environmental worthiness, data distribution measurement, performance data modeling, meta-dataset

中图分类号:

TB114

张书锋, 宋国锋, 李兴格, 陈循. 含AI模型机电系统环境适应性试验与评估方法[J]. 机械工程学报, 2025, 61(16): 358-373.

ZHANG Shufeng, SONG Guofeng, LI Xingge, CHEN Xun. Environmental Worthiness Testing and Evaluation of Electromechanical System Containing AI Model[J]. Journal of Mechanical Engineering, 2025, 61(16): 358-373.

参考文献

[1] 中国人民解放军总装备部电子信息基础部. 装备环境工程通用要求: GJB 4239—2001[S]. 北京:总装备部军标出版发行部, 2001. The Electronic Information Foundation Department of the General Armament Department of PLA. General requirements for materiel environmental engineering: GJB 4239—2001[S]. Beijing: Military Standard Publishing and Distribution Department of the General Armament Department, 2001.
[2] 中国人民解放军总装备部电子信息基础部. 装备环境工程术语: GJB 6117—2007[S]. 北京:总装备部军标出版发行部, 2007. The Electronic Information Foundation Department of the General Armament Department of PLA. Materiel environmental engineering terms: GJB 6117—2007[S]. Beijing: Military Standard Publishing and Distribution Department of the General Armament Department, 2007.
[3] 韦正现. 智能装备试验与测试的挑战与对策思考[J]. 测控技术, 2021, 40(2): 1-5. WEI Zhengxian. Challenge and countermeasure of intelligent equipment experiment and test[J]. Measurement & Control Technology, 2021, 40(2): 1-5.
[4] 李颖,王荣煊,万成麟,等. 分体式飞行汽车立体环境感知系统设计及试验研究[J]. 机械工程学报, 2024, 60(10): 102-111. LI Ying, WANG Rongxuan, WAN Chenglin, et al. Design and experiment research of 3D environmental perception system for split-type flying vehicle[J]. Journal of Mechanical Engineering, 2024, 60(10): 102-111.
[5] 蒋瑜. 频谱可控的超高斯随机振动环境模拟技术及其应用研究[D]. 长沙:国防科学技术大学, 2005. JIANG Yu. Research on the simulation of super-gaussian random vibration environment with controllable frequency spectrum and its applications[D]. Changsha: National University of Defense Technology, 2005.
[6] LI Li, WANG Xiao, WANG Kunfeng, et al. Parallel testing of vehicle intelligence via virtual-real interaction[J]. Science Robotics, 2019, 4(28): eaaw4106.
[7] 朱冰,张培兴,赵健,等. 基于场景的自动驾驶汽车虚拟测试研究进展[J]. 中国公路学报, 2019, 32(6): 1-19. ZHU Bing, ZHANG Peixin, ZHAO Jian, et al. Review of scenario-based virtual validation methods for automated vehicles[J]. China Journal of Highway and Transport, 2019, 32(6): 1-19.
[8] CARLSON J, MURPHY R R. How UGVs physically fail in the field[J]. IEEE Transactions on Robotics, 2005, 21(3): 423-437.
[9] 邵禹铭. 基于虚拟现实的无人车测试环境构建研究[D]. 西安:长安大学, 2019. SHAO Yuming. Research on the construction of unmanned vehicle test environment based on virtual reality[D]. Xi’an: Chang’an University, 2019.
[10] 王田苗,张韬懿,梁建宏,等. 基于再生能源的极地漫游机器人研究及现场试验[J]. 机械工程学报, 2013, 49(19): 21-30. WANG Tianmiao, ZHANG Taoyi, LIANG Jianhong, et al. Design and field test of a rover robot for antarctic based on renewable energy[J]. Journal of Mechanical Engineering, 2013, 49(19): 21-30.
[11] LI Nianshi, LIU Xiaoyong, YU Bendong, et al. Study on the environmental adaptability of lithium-ion battery powered UAV under extreme temperature conditions[J]. Energy, 2021, 219: 119481.
[12] OMIDSHAFIEI S, AGHA-MOHAMMADI A A, CHEN Y F, et al. Measurable augmented reality for prototyping cyberphysical systems: A robotics platform to aid the hardware prototyping and performance testing of algorithms[J]. IEEE Control Systems Magazine, 2016, 36(6): 65-87.
[13] 武月琴,傅耘. 装备环境适应性表征方法研究[J]. 装备环境工程, 2008, 5(6): 52-55. WU Yueqin, FU Yun. Research on characterization of equipment environmental worthiness[J]. Equipment Environmental Engineering, 2008, 5(6): 52-55.
[14] 王群,陈之光,汪文. 基于灰色理论的电子装备环境适应性评判模型[J]. 装备环境工程, 2010, 7(6): 94-98. WANG Qun, CHEN Zhiguang, WANG Wen, et al. Environmental worthiness model of electronic equipment based on grey theory[J]. Equipment Environmental Engineering, 2010, 7(6): 94-98.
[15] 张礼学,王中伟. 平流层飞艇环境适应性评价模型[J]. 航空学报, 2013, 34(4): 719-726. ZHANG Lixue , WANG Zhongwei. Environmental adaptability evaluation model for stratospheric airships[J]. Acta Aeronautica et Astronautica Sinica, 2013, 34(4): 719-726.
[16] 孟光磊,李树发,刘彬斌,等. 防空预警雷达高原环境适应性评估的自学习模糊灰度方法[J]. 兵工学报, 2022, 43(1): 98-110. MENG Guanglei , LI Shufa , LIU Binbin , et al. Self-learning fuzzy grey method for plateau environmental adaptability assessment of air defense early-warning radar[J]. Acta Armamentarii, 2022, 43(1): 98-110.
[17] HONG Y, LIAN J, XU L, et al. Statistical perspectives on reliability of artificial intelligence systems[J]. Quality Engineering, 2023, 35 (1): 56-78.
[18] DENG W, GOULD S, ZHENG L. What does rotation prediction tell us about classifier accuracy under varying testing environments?[C]//Proceedings of the 38th International Conference on Machine Learning. New York: PMLR, 2021: 2579-2589.
[19] MIN J, HONG Y, KING C B, et al. Reliability analysis of artificial intelligence systems using recurrent events data from autonomous vehicles[J]. Journal of the Royal Statistical Society Series C: Applied Statistics. 2022, 71 (4): 987-1013.
[20] JAFARZADEH M, AHMAD T, DHAMIJA A R, et al. Automatic open-world reliability assessment[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, Waikoloa, HI, USA. NewYork: IEEE, 2021: 1984-1993.
[21] 彭湃,耿可可,王子威,等. 智能汽车环境感知方法综述[J]. 机械工程学报, 2023, 59(20): 281-303. PENG Pai, GENG Keke, WANG Ziwei, et al. Review on environmental perception methods of autonomous vehicles[J]. Journal of Mechanical Engineering, 2023, 59(20): 281-303.
[22] 秦逸. 面向环境不确定性的 CPS 测试技术研究[D]. 南京:南京大学, 2018. QIN Yi. Software testing for cyber-physical systems suffering uncertainty[D]. Nanjing: Nanjing University, 2018.
[23] OBUKHOV A, KRASNYANSKIY M. Quality assessment method for GAN based on modified metrics inception score and Fréchet inception distance[J]. Proceedings of 4th Computational Methods in Systems and Software, 2020, 1: 102-114.
[24] HEUSEL M, RAMSAUER H, UNTERTHINER T, et al. GANs trained by a two time-scale update rule converge to a local nash equilibrium[J]. Asian Journal of Applied Science and Engineering, 2019, 8: 25-34.
[25] 贾学志,张雷,安源,等. 空间光学遥感器精密调焦机构设计与试验[J]. 机械工程学报, 2016, 52(13): 25-30. JIA Xuezhi, ZHANG Lei, AN Yuan, et al. Design and experiment research on precision focusing mechanism of space remote sensor[J]. Journal of Mechanical Engineering, 2016, 52(13): 25-30.
[26] 向红标,杨大虎,杨璐,等. 复杂环境下磁弹性微型游泳机器人的路径规划与识别跟踪[J]. 机械工程学报,2023, 59(5): 89-99. XIANG Hongbiao, YANG Dahu, YANG Lu, et al. Path planning and recognition tracking of a magnetoelastic miniature swimmer in complex environment[J]. Journal of Mechanical Engineering, 2023, 59(5): 89-99.
[27] LI Xingge, ZHANG Shufeng , CHEN Xun , et al. Robustness of visual perception system in progressive challenging weather scenarios[J]. Engineering Applications of Artificial Intelligence. 2023, 119: 105740.

含AI模型机电系统环境适应性试验与评估方法

Environmental Worthiness Testing and Evaluation of Electromechanical System Containing AI Model

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics

本文评价

[1]	石佳东, 马国政, 李国禄, 李振, 赵海朝, 王海斗. 典型工况环境中类金刚石薄膜摩擦学性能研究进展[J]. 机械工程学报, 2025, 61(10): 19-35.
[2]	王田苗;张韬懿;梁建宏;陈蛟. 基于再生能源的极地漫游机器人研究及现场试验[J]. , 2013, 49(19): 21-30.